Insect Identification Project for Agricultural Advancement

Through the use of image data, we developed an artificial intelligence system which is able to predict an insect’s species based on a photo of a given insect.

Insects are complex creatures whose influence and characteristics can drive ecosystems or destroy plant populations. Therefore, it is imperative to recognize the differences between them and whether they pose harm or benefits, yet the majority of humans do not have a keen view for the differences in insect qualities. Through the use of image data, we developed an artificial intelligence system which is able to predict an insect’s species based on a photo of a given insect. To assess the image data, we developed a convolutional neural network (CNN) which uses a series of neurons to output a result which we can interpret to identify the given insect. The highest validation accuracy we received was approximately 80%, which given the very subtle differences between the insects we used, the amount of categories we were defining from, and the limited number of images in each category, was a decently high validation accuracy. From this, we could conclude that we can not only identify a species, but with relatively high accuracy, giving us insight into how insect identification is a valid and applicable process which can be implemented into agricultural and pest control technology.

Explore More!

Research Paper

Source Code